Senior ML Engineer & AI Researcher — 15+ years building production AI systems, from on-device computer vision to LLM compression and multimodal intelligence.
I'm the Founder & CEO of OM Enterprises — AI Research & Engineering Labs, focused on cutting-edge ML research and enterprise AI services measured against real KPIs. My mission: build the next generation of AI as ubiquitous, interactive intelligence that runs on multimodal data — audio, text and video — with high throughput and low latency.
On a typical day I counsel organizations alongside AI researchers by posing hard questions, critiquing work and unblocking problems. The rest goes to hosting and presenting at study groups, contributing to open source, and competing on Kaggle.
My research interests cover anything that moves, and anything on the audible spectrum. Lately that means State Space Models, multimodal models across audio, text, image and video including Vision Language Models, Transformers and their derivatives, Reinforcement Learning, Graph Neural Networks and multilingual LLMs — and how they solve self-driving autonomous systems: vehicles, drones and robotics.
Over 15+ years I've contributed to 50+ research papers, advising researchers and PhDs on engineering optimizations and bouncing novel ideas to scale research across ICLR, ICML, CVPR, NeurIPS, ECCV and ML Summit. I also mentor, advise and sometimes judge founders at a startup accelerator on efficient engineering, product-market fit and funding.
Applied work spans retail, content creation, media (podcasts, news and publishing), healthcare, gaming (storytelling through immersive voices), avatars and voice agents with localized accents, logistics, recruiting, and sales and support. I've deployed on-device (iOS), on edge (DeepStream SDK) and at scale (CUDA/Triton).
I'm a TWiML Contributor Lead (5+ years) running the weekly Generative AI Problem Solving, Research & Impact meetup, a Cohere Labs Community Lead & Researcher, an MLT Tokyo contributor, and a core organizer of Lucknow AI Labs.
Out of 1,427 teams (2024)
CLIP-RSICD fine-tuning
Rank 1,532 / 202K
NER, Text Gen, Image Gen
Advised across top ML venues
Enterprise systems to AI research
Production-grade ML systems with reproducible pipelines, W&B instrumentation, and benchmark results.
Three years of competing in the ARC Prize — the benchmark for measuring AI's ability to learn new tasks from few examples.
1 Silver + 3 Bronze medals across 21 completed competitions spanning CV, NLP, medical AI, and reasoning.
Open to Senior ML Engineer, Research Engineer, CV Engineer, and Senior SDE roles.